基于MindSpore的农作物病虫害识别系统设计  

Design of crop pest identification system based on MindSpore

作　　者：杨祖鉴 黄勇萍[1] Yang Zujian;Huang Yongping(College of Mathematics,Physics and Electronic Information Engineering,Guangxi Normal University for Nationalities,Chongzuo 532200,China)

机构地区：[1]广西民族师范学院数理与电子信息工程学院,广西崇左532200

出　　处：《无线互联科技》2023年第3期59-61,共3页Wireless Internet Technology

基　　金：2021年国家级大学生创新创业训练计划项目,项目名称:基于MindSpore的农作物病虫害识别系统的研究与设计,项目编号:202110604008;广西高校中青年教师科研基础能力提升项目,项目名称:基于RFID与WiFi的跨频农田环境监测技术研究,项目编号:2020KY20017。

摘　　要：广西是农业大省区,农作物种类繁多,包括甘蔗、水稻等。部分地区以农作物为主要经济来源,由于气候原因,农作物病虫害频发。文章研究了广西主要农作物的常见病虫害,采集了甘蔗等农作物的常见病虫害数据,使用AI计算框架MindSpore以及深度学习算法MobileNetV2实现了数据的特征提取以及模型训练,使用端侧推理框架MindSpore Lite实现了快速准确的端侧推理。测试结果表明,该系统有较高的识别准确率,并且与传统的识别软件相比,系统可以进行无网络、轻量级实时识别。Guangxi is a large agricultural province,a wide variety of crops,such as sugar cane,rice and so on.In some areas,crops are the main source of economy,and crop diseases and insect pests occur frequently due to climate reasons.This paper studied the common diseases and insect pests of major crops in Guangxi,collected the data of common diseases and insect pests of sugarcane and other crops,used the AI computing framework to achieve the feature extraction and model training of the data using MindSpore and the deep learning algorithm MobileNetV2,and used the end-side inference framework to achieve fast and accurate end-side inference framework MindSpore Lite.The test results show that the proposed system has a high recognition accuracy,and compared with the traditional recognition software,the system can perform network-free,lightweight real-time recognition.

关 键 词：MindSpore 病虫害识别 端侧推理 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TP311[自动化与计算机技术—计算机科学与技术]